An ecological study on childhood autism

Table 1 Descriptive statistics for variables included in our regression analysis and results of the univariate regression analysis for each variable

Variable	Mean	Standard deviation	25% quartile	75% quartile	Regression Coefficient	t- statistic	p-value
State Name
% autism in Washington counties (n=35)	0.24	0.13				Reference state
% autism in California counties (n=57)	0.18	0.07			−0.149	−9.32	<0.001
% autism in Oregon counties (n=28)	0.68	0.25			0.291	13.17	<0.001
Drinking water derived from surface sources (%)	48.51	34.70	15.0	79.75	0.073	4.09	<0.001
Average unemployment (%)	7.43	3.03	5.21	8.96	−0.025	−4.28	<0.001
Suicide rate (per 100,000)	14.24	5.09	10.13	17.10	0.074	4.17	<0.001
Pollutants
EPA risk of neurological disease from air emissions (hazard index )	0.059	0.054	0.032	0.061	0.071	3.95	<0.001
Crop density (Square km of land used to grow crops per county square km )	0.15	0.18	0.0247	0.21	−0.019	−0.97	0.334
Population (number of people per square km)	361	1563	19	163	−0.008	−0.41	0.685
Meteorological parameters
Average annual precipitation (cm)	108.87	70.97	54.97	165.71	0.092	5.37	<0.001
Median elevation (ft)	440.3	479.1	146.0	516.8	−0.008	−0.42	0.676
Heating degree days (⁰C*days)	2933.3	1045.7	1914.0	3718.4	0.030	1.58	0.118

The outcome variable used in the analyses was the square root transformed percent of autism for each county and the regression coefficients are for standardized variables. Regressions were run using data from 120 counties in Washington, Oregon, and California.

ISSN: 1476-072X